Regression Help for Masters thesis


TS Contributor
Overall set of questions/hypotheses? Give us some idea of the starting point.

Type of predictors (categorical, continuous)? Time extent? Random effects?


In addition to what jpkelly requested: Format of data? Is it in Excel? Familiarity with statistics computer? programs? Access to stats programs (like SPSS)?


TS Contributor
OK, since you have a time series, you have temporal autocorrelation. This should probably be account for somehow. Is this annual data, monthly data? Do you want to estimate your effects overall (independent of year), by year, by month? Do you have time only for emerging markets, or do you also have other data?

I would start by plotting all the data (two time series) and then sitting down and writing down all the questions you have about the data, as the perceived patterns relate to your ideas about the economy. Try to imagine everything you will want to test with the time series. After that, we can help suggest tests.

So, seriously, plot the data and post here. That will get the discussion started.

(A side note: If you're just starting with statistics, I might recommend NOT using SPSS. I've found that the ability to simply click on a statistic makes you try a bunch of tests without thinking about your data. I'm biased, but I would suggest R as a starting point. If you grow comfortable with that, you can then choose to use SPSS or other GUI-based programs. With R, I've found that I think critically about my data and don't waste lots of time trying different --that is, inappropriate--tests.)


Ninja say what!?!
I'm just a little curious here, and please don't take this offensively. Are you really planning on having strangers on a forum walk you through something as important as your master's thesis? Though we can definitely give you advice, I sure hope we aren't your only resource.

In addition to what the other members have asked already, here are some more questions. Who have you talked to already? What have they told you? How much experience do you have with data analysis?